The typical Premium Bonds holder wins less than the prize fund rate suggests — the median year, modelled exactly, beside the advertised average.
The prize fund rate is an average, and averages mislead here by design: rare large prizes pull the mean up, so the TYPICAL holder wins less — often nothing. On £1,000 of Bonds at an example 4% fund rate and 1-in-22,000 odds, the advertised average is £40 a year — while the median holder wins £0: the chance of winning nothing at all across the whole year is 58%. At the calculator's £5,000 default the gap stays wide — £200 advertised against a median floor of £75. The calculator models the median exactly (a Poisson count of wins, each valued at the minimum prize as a stated floor) and takes the rate and odds as your inputs, because NS&I varies both and publishes the current pair.
Your holding, and the prize fund rate and odds as NS&I currently publishes them — both are examples here until you enter today's pair.
The typical holder wins less than the prize fund rate suggests, because the average is pulled up by large prizes most holders never see. The headline above counts each median win at the £25 minimum prize — a floor, stated as one: the median COUNT is exact, but NS&I varies how prize money is split across values, so no estimate of a typical year's value beyond the floor is printed here.
Prizes are tax-free. The rate and the odds here are yours to enter — both are variable and NS&I publishes the current pair; nothing on this page asserts either. A planning estimate, not financial advice.
| Tax/source year | 2026/27 |
|---|---|
| Region | UK-wide: savings interest is taxed on the UK bands and thresholds in every UK nation, including for Scottish taxpayers |
| Account rates | Never assumed. Any interest rate in a savings calculation is an input you set, stated beside the result. |
| Your other income | Wages and pension income decide how much Personal Allowance and starting rate are left for interest, and which Personal Savings Allowance band applies. |
|---|---|
| The interest itself | Interest counts towards the income that sets your band, so a large interest year can shrink the allowance that was meant to cover it. |
| The account's wrapper | The same balance at the same rate is taxed outside an ISA and untouched inside one; the wrapper, not the rate, decides whether the allowances are needed. |
| What is not decided here | PayBreakdown does not decide payroll correctness, lender approval, benefit entitlement or employer compliance, and does not give regulated financial, mortgage, debt, tax, payroll or legal advice. |
|---|
The prize fund pays out across values from the minimum up to the top prize, and the biggest prizes belong, in any given year, to almost nobody. The average spreads every prize across every holder; the median asks what the person in the middle actually receives. At a large holding the two are still far apart: £50,000 at the example inputs averages £2,000, while the median year's wins, counted at the minimum prize, floor at £675 — the gap between the two figures is everything paid above the minimum, in a split NS&I varies and this page deliberately does not estimate. At a small holding the average stops describing anyone: the median £1,000 holder's year is zero prizes.
Each £1 Bond enters the monthly draw with the odds you supply, so a year's prize count follows a well-defined distribution, and the median of that count is computed exactly — no simulation, no hand-waving. Each median win is then valued at the £25 minimum prize. That makes the printed figure a floor on the median year, and it is stated as one — nothing more. How the prize money splits across values above the minimum is NS&I's to vary, and it does vary it, so a typical year's actual value sits at or above the floor by a margin this page deliberately does not estimate — exactly at it, when the median count is zero. Prizes are tax-free, which is the honest half of the pitch: for a saver whose interest has already outgrown their Personal Savings Allowance, a tax-free median compares against an ordinary account's after-tax interest, and gains relative value exactly there.
It never asserts the prize fund rate or the odds. NS&I varies both — sometimes several times a year — and a page that printed them would be wrong within months while looking authoritative the whole time. Both are inputs, the current pair is on NS&I's Premium Bonds page, and the source below is that page. Whether Premium Bonds beat a savings account for you depends on your allowances, your tax rate and your appetite for a zero year: the savings interest tax calculator works the taxed side of that comparison.
Change any figure above and the result moves with it. A payslip can differ from any calculator because payroll works pay period by pay period, so a tax-code change, a bonus, arrears or a refund can land in one month and not the next. See how each deduction is worked out.
No — it is a true average and NS&I states it as one. The trap is reading an average as an expectation for yourself: the distribution is so skewed that most holders receive less than the rate suggests in most years, which is what the median row shows.
Because that makes the figure a floor with a clear meaning, rather than a point estimate resting on the full prize table — which NS&I reshapes as the fund rate moves, and which this page would then misstate within months. The median count is exact; the value of a typical year sits at or above the floor by however much of the fund is allocated to mid-size prizes at the time, which is a published, changing fact this page points you to rather than freezes.
No — prizes are tax-free. That matters most to savers whose interest has outgrown their allowances: a tax-free median compares against an ordinary account's AFTER-tax interest, which is the comparison the savings interest tax calculator runs.
Last updated 2026-08-28. Checked by Sean Elsmore. Rates last checked against GOV.UK on 2026-08-07. Estimates are for planning and should be checked against official records where the decision matters.